https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81693
--- Comment #29 from H.J. Lu <hjl.tools at gmail dot com> --- (In reply to Eric Gallager from comment #28) > (In reply to H.J. Lu from comment #27) > > By definition of the "naked" attribute, the program is responsible > > to manage stack. Since simulated interrupt functions don't follow > > the normal software calling convention and there is no attempt > > made to accommodate it in the naked function, the outgoing stack > > is properly aligned to requirement of software calling convention. > > Since Darwin checks the incoming stack alignment, the failure is > > expected. > > So XFAIL it and close this bug then? Yes.